frames
frames
#
posed_frame — one decoded frame written the way every dataset builder writes it.
A dataset builder decodes a frame from its source (a keyframe, a video, a
panorama sweep) and hands it here to become a PosedImage: shrunk so its
longest side is at most max_dimension (never enlarged), re-encoded as a
JPEG in the recipe workspace, with the pose's intrinsics re-expressed at the
written size so every pixel stays a ray.
posed_frame(picture, pose, depth, max_dimension)
#
picture as a posed image at most max_dimension pixels along its longest side.
Parameters:
| Name | Type | Description | Default |
|---|---|---|---|
picture
|
Image
|
The decoded frame. |
required |
pose
|
Pose
|
The camera, its intrinsics expressed at |
required |
depth
|
Depth
|
The depth the frame measured, |
required |
max_dimension
|
int
|
The longest image side, in pixels, as written. |
required |
Raises:
| Type | Description |
|---|---|
ValueError
|
If |